It seems my earlier tweet about MSFT and FP communities has caused some confusion so let's expand the topic a bit.

Let me go through some of my anecdotal experience I have made while speaking at multiple FP conferences for last 4 years 1/n

#fsharp #devrel #functionalprogramming https://twitter.com/k_cieslak/status/1199846544587055104

WARNING: This will be long thread

The entry point of the discussion is the observation - if you use FP language for "real projects" you're more likely to use AWS over Azure. In fact, Azure is probably not even a thing you consider. And there are multiple reasons for that.

2/n

<side note>

Before going further - F# is an outlier here. Since it's a .Net language, owned by MSFT the F# user base is traditionally using a lot of MSFT platforms. So in this thread I want to focus mostly on different FP communities

3/n

However, at the same time, one could argue that, for example, it's easier to use F# on AWS Lambda rather than on Azure Functions... which is great example of MSFT attitude towards FP languages...

</side note>

4/n

First of all, many of those communities has terribly outdated view of MSFT - things that are obvious for us, .Net people, like "new MSFT that loves OSS and community", ".Net Core being OSS and xplat" are not something well known outside our bubble

5/n

Why it happens? Mostly because no one from MSFT even tries to "talk" with those people - I've attended multiple FP events over last few years and I can count on one hand number of people from MSFT present at such conferences.

6/n

And if anyone is around those are people from "F# team" - like @_cartermp or @dsymetweets - and DevRel ain't their main job, which means those efforts are naturally limited (in terms of number of attended conferences etc)

7/n

The second group of people doing any "positive marketing" at FP events are folks from F# community - and those efforts are also naturally limited and lack a bit of legitimacy (it's hard to convince people MSFT cares about them if no one ever from MSFT talked with them)

8/n

But we do have some success stories here, where strong F# presence on conference clearly impacted the community (more and more people attending F# talks every year, less and less questions like "So .Net runs only on Windows, right?")

9/n

The second part of the issue is more on MSFT side - FP is not mentioned at MSFT conferences (even F# is ignored at conferences like Build), there are no blog posts about FP from MSFT, there is no documentation.

10/n

<side note>

I talk a lot about serverless programming in this thread, mostly because I like the idea, and I strongly believe it's perfect programming model for using FP languages in the cloud

</side note>

11/n

There is also lack of any library support for running FP languages on Azure. Why should I deploy Haskell to Azure Functions (with some undocumented, not supported hacks like creating PowerShell Function and then execute Haskell under the hood)?

12/n

Instead, I can go to AWS, drop library, and get direct Haskell support with nice API - https://theam.github.io/aws-lambda-haskell-runtime/

Or OCaml - https://github.com/anmonteiro/aws-lambda-ocaml-runtime

Searching for Ocaml Azure Functions yields 0 results.

13/n

And just to be clear - I'm not saying MSFT should create those libraries, and documentation. But it's all related - because MSFT ignores FP communities, no one even consider Azure as a choice, so community doesn't create them on their own, so "normal" user won't pick Azure

14/n

In the end one of the biggest values of DevRel in this context is creating userbase of active contributors that will do stuff as OSS projects that you as a company don't want to do. MSFT is totally failing at that in case of FP communities

15/n

So the last questions is... why MSFT should even care? I mean, they're winning in $$$ race by targeting big communities.

And this is undeniable fact - all FP communities together are probably less than rounding error in measurements of numbers of C# developers ;-)

16/n

At the same time - FP languages are gaining momentum. We can observe constant user growth. We can observe the success stories - and it's not always just about plain users of numbers. Single success story using FP language may be worth all investment in FP + Azure DevRel

17/n

For example Jet - F# startup - is spending quite a lot of money ( ;-) ) on Azure every month. Or imagine next WhatsApp (originally created in Erlang) being hosted on Azure?

18/n

It's also the question of mind-share and different perspective on software development. FP communities (and any niche communities) tend to have quite a high percentage of people that are active in OSS communities, really passionate about what they do, having different ideas

19/n

So why not convince them to use their brain power, their different ideas to make everyone life better? We cannot create thriving ecosystem when everyone involved has single background and similar opinions on software development.

20/n

Lastly, FP seems to be just good fit for cloud programming in general - from stateless web servers that can be easily scaled, through serveless programming which is naturally fitting the paradigm, to actor model that was pioneered in FP world.

21/n

So maybe it's time for MSFT to push frontier in this area? The funny thing is that Amazon is not doing a lot in promotion their products in FP communities neither, but they have head-start thanks to terrible reputation of old MSFT. Which could be changed if MSFT tried...

22/n

So summing up:

* MSFT is missing the boat with FP + Cloud
* Lot of issues are not technical, but related to lack of knowledge about "new MSFT"
* MSFT is not doing enough to fix that problem
* And they should to make better platform for everyone

23/24

The whole thread was based on my anecdotal experience, take it with huge pinch of salt, your millage my vary, I probably just hate MSFT and randomly rant because I'm bad person ;]

Thanks for reading.

24/24